Allow caching of `hashCode` on Freezed models

[Matthiee]

Is your feature request related to a problem? Please describe.
During the profiling of the app I work on, I noticed that hashCode call for larger, deeply nested Freezed models is rather expensive. Upon further inspection, the hashCode is recomputed every time it is accessed.

Describe the solution you'd like
Repeated calls to .hashCode should return the cached value instead of recomputing.

Additional context
A similar request was made in built_value google/built_value.dart#760

Currently Freezed generates

@override
int get hashCode => Object.hash(runtimeType,propertyA,const DeepCollectionEquality().hash(someCollection),...

With cached hashCode

int? _cachedHashCode;

@override
int get hashCode => _cachedHashCode ??= Object.hash(runtimeType,propertyA,const DeepCollectionEquality().hash(someCollection),...

[rrousselGit]

I don't think there's much value here. You can do it yourself very easily:

mixin CachedHash {
  @override
  late final hashCode = super.hashCode;
}

I'm not entirely against the idea. But it seems like we're adding a new API here for something that can be down using a very succinct syntax already.

[Matthiee]

Some of the reasons I think there is still value to be had in this:

• By giving Freezed control of hashCode implementation, it should generate the best possible version
• Unless one manually reviews the generated files or uses profiling tools, one might not know there is an almost free performance boost to be had in specific situations
• I have 400+ different Freezed models, I now have to override the hashCode manually for all of them
• For all of my models, I now have to manually reason whether or not they would benefit from a cached hashCode

I'm happy to help create a PR for this feature.